Evaluation and bias corrections of gridded precipitation data for hydrologic modelling support in Kabini River basin, India

A comprehensive assessment and bias corrections of two gridded daily precipitation products, based on gauge-only and multi-satellite observations, are undertaken in this study using independent rain gauge data in and around the Kabini River (KR) basin in South India. The KR basin, with complex terrain, highly variable precipitation and heterogeneous land use poses challenges in the development of accurate gridded precipitation products. The gauge-only gridded precipitation data from the India Meteorological Department (IMD) and multi-satellite gridded precipitation product from the Tropical Rainfall Measuring Mission (TRMM) Multi-satellite Precipitation Analysis (TMPA)-3B42 are evaluated at daily and monthly scales in this study. Both gridded precipitation products available at 0.25 degrees spatial resolution are assessed using independent 57 rain gauge observations for the period of 2009-2016. Multiple visual and statistical metrics have been utilized to assess the error characteristics and capability of these gridded precipitation estimates in replicating extremes. Results indicate that gauge-only gridded product is generally better than the multi-satellite precipitation product. The multi-satellite product notably overestimates light precipitation and underestimates extreme precipitation over the study region. To mitigate the overestimation of dry days in the TMPA-3B42 estimates, a dry-day correction method is developed and is applied using nearby rain gauge observations. Furthermore, a quantile-based correction is also applied to both gridded precipitation products after confirming the stationarity of the data, which substantially improved both precipitation estimates for distributed hydrological modelling studies in the KR basin.